Article Names from Top 5000 should be validated

[Lichtschalter-5000]

Snapshot of the Top 5000 in the Wayback machine
(It's on rank 2454)

Because the article name was too long, it has been shortened. The result is no longer the name of an article.
The list of page names is built by retrieving the text content of the links (WikiUpdater.js:92). Instead, the content of the href-attribute could be stored (although this raises the need of more refactoring, because the previous "Albert Einstein" would now be "Albert_Einstein").
Another option would be to check whether an article really exists.

Maybe this is connected to #159, unlikely though, because the backlinks there wouldn't have worked either.